In recent years, high-efficiency transmission by digitalization has been widely adopted. When a multi-phase modulation scheme is adopted in wireless communication as a method of achieving high-efficiency transmission, a technique of suppressing nonlinear distortion by linearizing input/output characteristics particularly of a power amplifier on a transmission side, and of reducing leakage power of an adjacent channel is important. Moreover, when an amplifier with poor linearity is used to improve power efficiency, it is desirable that nonlinear distortion be compensated.
In mobile communication such as wideband code division multiple access (W-CDMA), transmission power of a transmission device is 10 milliwatts (mW) to several tens of watts (W), the input/output characteristics of the power amplifier has nonlinearity. In this case, the input/output characteristics of the power amplifier is expressed by a distortion function f(p). This nonlinearity causes nonlinear distortion, and a frequency spectrum around a transmission frequency f(0) leaks to an adjacent channel due to side lobe, to result in causing adjacent interference. That is, because of the nonlinear distortion, the power of transmission wave that leaks to an adjacent frequency channel increases. The amount of this leakage power is indicated by an adjacent channel power ratio (ACPR) that is a ratio between power of a channel of an own device having f(0) in center and adjacent leakage power that leaks to an adjacent channel. That is, it can be understood that when the adjacent leakage power becomes higher than the power of a channel of an own device, the leakage power has increased. This leakage power is a noise to other channels and deteriorates the communication quality of other channels that are influenced by the leakage power.
Leakage power is small in a linear region and large in a nonlinear region in input/output characteristics of a power amplifier, for example. Therefore, to increase output of the power amplifier, widening the linear region in the input/output characteristics of the power amplifier can be considered. However, to widen the linear region, using a high-performance amplifier is considered, and it results in increase in cost and the size of an apparatus.
Therefore, to suppress distortion of a transmission signal that has passed through an amplifier, a distortion compensation apparatus in a digital linear-distortion compensation scheme is arranged in a wireless communication apparatus in some cases. The digital linear compensation scheme is a scheme in which a carrier wave that is obtained by performing quadrature modulation by a modulation signal is used for feedback detection, and amplitudes of the modulation signal (transmission baseband signal) and a feedback signal (feedback baseband signal) are digitized to be compared, and a distortion compensation coefficient is sequentially updated based on a result of comparison. For example, a distortion compensation apparatus in the digital linear-distortion compensation scheme acquires inverse characteristics of amplification characteristics using a distortion compensation coefficient. The distortion compensation apparatus in the digital linear-distortion compensation scheme then adds distortion of the inverse characteristics of amplification characteristics to a transmission signal before being input to an amplifier. By addition of the distortion of the inverse characteristics, distortion of the transmission signal that has passed through the amplifier is suppressed, and the nonlinearity of the amplifier is compensated.
Among distortion compensation apparatuses as described, there is, for example, one that stores distortion compensation coefficients used for distortion compensation in a lookup table (LUT), and reads a compensation coefficient from the LUT by specifying an address corresponding to the power level of a transmission signal. In such a distortion compensation apparatus, for example, LUT addresses to compensate the nonlinearity of the amplifier according to the power has been generated using, for example, power, amplitude, and a value obtained by performing log conversion on power.
Moreover, a technique in which distortion compensation is performed only when an amplitude value of an input modulation signal is within a predetermined range has been known. Furthermore, a technique in which frequency of specification of each address of LUT is monitored and an address proportional to the power level is converted into an actual LUT address such that the specification frequency is equalized has been known. A technique in which the power level of a signal is detected using a multiplier and lookup elements to an encoding signal that is generated from a square signal of a reception signal, and closed loop gain control is performed using a detected power level has been known.    Patent Document 1: Japanese Laid-open Patent Publication No. 2007-208684    Patent Document 2: Japanese Laid-open Patent Publication No. 2007-49251    Patent Document 3: Japanese Laid-open Patent Publication No. 2005-184847
However, because an allocation method of address is simple in a conventional method of generating an LUT using power, amplitude, or a value obtained by log conversion, an optimal LUT address according to a frequency distribution of input signals cannot be generated. Therefore, by the conventional method, it is difficult to reproduce the nonlinearity of an amplifier according to the power, and the distortion compensation performance can be degraded.
Even if the technique in which distortion compensation is performed only when an amplitude value of an input modulation signal is within a predetermined range is applied, it is difficult to reproduce the nonlinearity of an amplifier according to the power, and the distortion compensation performance can be degraded. Furthermore, even by the technique in which an address proportional to the power level is converted into an actual LUT address such that the designation frequency is equalized is applied, if the circuit size is small, it is difficult to reproduce the nonlinearity of an amplifier according to the power, and the distortion compensation performance can be degraded. Moreover, even if the technique in which closed loop gain control is performed using a power level detected using a squared signal of a reception signal is applied for generation of LUT, it is still difficult to reproduce the nonlinearity of an amplitude according to the power.